# TPU 8i Max

Google has added two systems to its eighth-generation TPU line: TPU 8i Max, aimed at very long context windows and the highest compute density per rack, and TPU 8tx, aimed at secure and confidential inference. The same generation already included TPU 8t for large-scale training and TPU 8i for real-time reasoning, which Google says deliver up to 2.7 times better training performance and 80% better inference price performance than the previous generation. The new systems keep the modular custom-silicon plus open-software-stack approach, rebuilding compute, networking and storage around the chip to ship specialized supercomputers in months rather than years.

Google has added two systems to its eighth-generation TPU line: TPU 8i Max, aimed at very long context windows and the highest compute density per rack, and TPU 8tx, aimed at secure and confidential inference. The same generation already included TPU 8t for large-scale training and TPU 8i for real-time reasoning, which Google says deliver up to 2.7 times better training performance and 80% better inference price performance than the previous generation. The new systems keep the modular custom-silicon plus open-software-stack approach, rebuilding compute, networking and storage around the chip to ship specialized supercomputers in months rather than years.